Skip to main content
Multi-turn conversations enable natural, context-aware interactions with your AI Voice Agents.

​
What are Multi-turn Conversations?

Multi-turn conversations are conversations where:
  • Context is maintained - Agent remembers previous turns
  • References work - “What about tomorrow?” refers to previous context
  • Natural flow - Feels like talking to a person

​
Example

# Turn 1
User: "What's the weather in London?"
Agent: "The weather in London is cloudy, 15°C."

# Turn 2 (agent remembers London from previous turn)
User: "What about tomorrow?"
Agent: "Tomorrow in London will be sunny, 18°C."

# Turn 3 (agent remembers the conversation)
User: "And the day after?"
Agent: "The day after tomorrow in London will be partly cloudy, 16°C."

​
How It Works

​
Context Maintenance

The agent automatically maintains context:
  1. Stores messages - All user and agent messages
  2. Includes tool results - Tool execution results in context
  3. Uses history - Full conversation history available to LLM
  4. Maintains session - Each session has isolated context

​
Turn Flow

Turn 1: User → Agent
    ↓ (context stored)
Turn 2: User → Agent (with context from Turn 1)
    ↓ (context updated)
Turn 3: User → Agent (with context from Turns 1 & 2)

​
Best Practices

​
Clear References

  • ✅ Be specific - “What about tomorrow?” works if context is clear
  • ✅ Provide context - “What’s the weather in London tomorrow?” is clearer
  • ✅ Use consistent terms - Use same terminology throughout

​
Context Management

  • Keep conversations focused - Long conversations may exceed limits
  • Use tools effectively - Tools help maintain context
  • Clear session boundaries - Start new sessions for new topics

​
Next Steps